太平洋大眼金枪鱼和黄鳍金枪鱼中原产地和种群连通性检测方案(激光剥蚀进样)

智能文字提取功能测试中

GISHERIESOCEANOGRAPHYFISHERIESOCEANOGRAPHYFish. Oceanogr. 25：3， 277-291，2016 278J. R. Rooker et al. Natal origin and population connectivity of bigeye andvellowfin tuna in the Pacific Ocean JAY R. ROOKER，1，2* R. J.DAVID WELLS，1，2DAVID G. ITANO，SIMON R. THORROLD4AND JESSICA M.LEE 'Department of Marine Biology， Texas ABM University atGalveston，1001 Texas Clipper Road， Galveston， TX 77554，U.S.A. Department of Wildlife and Fisheries Sciences， Texas AGMUniversity， College Station， TX 77843， U.S.A.National Marine Fisheries Service， Pacific Islands Region，1601 Kapiolani Boulevard， Suite 1110， Honolulu， HI 96814，U.S.A. *Biology Department， Woods Hole Oceanographic Institution，Woods Hole， MA 02543， U.S.A ABSTRACT Natural chemical markers (stable isotopes and traceelements) in otoliths of bigeye tuna (Thunnus obesus)and yellowfin tuna (T. albacares) were used to investi-gate their origin and spatial histories in the westernand central Pacific Ocean (WCPO). Otolith chem-istry of young-of the-year (YOY) T. obesus and T. al-bacares from four regions in the WCPO was firstdetermined and used to establish baseline chemicalsignatures for each region. Spatial variation in stableisotope ratios of YOY T. obesus and T. albacares wasdetected， with the most noticeable difference beingdepleted otolith818o values for both species from thefar west equatorial and west equatorial regions relativeto the central equatorial and Hawaii regions.Elemen-tal ratios in otoliths were also quantified for YOYT. obesus and T. albacares collected in 2008， and several showed promise for distinguishing YOY T. obesus(Mg：Ca，Mn：Ca， and Ba：Ca) and T. albacares (Li：Caand Sr：Ca). The natal origin of age-1 to age-2+T. obe-sus and T. albacares was then determined for tworegions of the WCPO， and mixed-stock analysis indi-cated that T. obesus and T. albacares in our west equa-torialsamplewerealmost entirelyfrom localproduction， with a minor contribution from centralequatorial waters. Similarly， T. albacares collected inHawaii were exclusively from local sources； however， a *Correspondence. e-mail： rookerj@tamug.edu ( Received 12 August 2 015 ) ( R evised v ersion accepted 20 January 2016 ) large fraction of T. obesus in Hawaii were classified tothe central equatorial region， suggesting that themovement of migrants from outside production zones(i.e.， south of Hawaii))are important to Hawaii'sdomestic fishery. Key words： equatorial IPacific， isotopes， migration，natal origin， otolith chemistry， pelagic， stable isotopes，trace elements， stock mixing INTRODUCTION Pelagic fishes that utilize open-ocean ecosystems arecapable of long-distance migrations that often crossinternational jurisdictions or ocean boundaries (Roo-ker et al.， 2008a， 2014； Block et al.， 2011). Unfortu-nately， migratory behaviors aswel1l as residencypatterns of many pelagic fishes are poorly understood，limiting our ability to effectively manage and conservecertain populations (Rooker et al.， 2014). This is par-ticularly true for tropical tunas， which are importantpredators in open-ocean ecosystems (Young et al.，2010； Ferriss and Essington， 2011)， and represent keycomponents of high seas fisheries in tropical waters(Joseph et al.， 2010). Fishing pressure and harvest sce-narios for tropical tunas commonly vary spatiallywithin an ocean basin， and， therefore， the populationdynamics of these stocks will be influenced by theirmovement patterns (Sibert and Hampton，2003). Several species of tropical tunas occur in the west-an (wepdern and central Pacific Ocean (WCPO)， and eventhough these tunas are capable of undertaking longdistance migrations (> 1000 km)， restricted move-ments and high site fidelity have also been reported(Hampton and Gunn， 1998； Itano and Holland， 2000；Schaefer et al.， 2011，2015). Defining natal origins，stock structure and migration routes of tropical tunasin the WCPO is critical for their management， partic-ularly for two species heavily targeted by fishing fleets：bigeye tuna(Thunnus obesus) and yellowfin tuna(T. albacares)(Schaefer， 2008). Results to date indi-cate that total and adult spawning stock biomass ofT. obesus andT. albacaressirintheWCPO havedeclined significantly over the past few decades withcatches now predicted to be close to or exceeding the maximum sustainableyyield((DIavies et al.， 2014；Harley et al.， 2014). These apparent population decli-nes are of particular concern because the stock struc-ture， migration patterns and the degree of mixingwithin the WCPO for both species are unresolved.Several approaches have been used to investigate themovement and mixing rates of T. obesus and T. albacares from different production areas in the WCPO，and some degree of stock or contingent structure maybe present in both species (Schaefer et al.， 2011，2015). Still， data are currently insufficient to moveaway from the premise of a ‘single'panmictic popula-tion for both T. obesus and T. albacares (Ward et al.，1997； Grewe and Hampton， 1998)， even though datafrom tagging and genetics research suggests that tropi-cal tunas display more site fidelity and potentially existas discrete subpopulations or management units withinthe WCPO (Hampton and Gunn， 1998； Itano andHolland， 2000； Sibert and Hampton， 2003； Greweet al.， 2015；Schaefer et al.，2015). In the present study， we examined the origin andmovement of young (age-1to age-2+) T. obesus andT. albacares in the WCPO using chemical markers(stable isotopes and trace elements) in their otoliths(ear stones). Similar to other highly mobile species，understanding movement patterns is of particularinterest because the potential for basin-scale migrationis high for young tunas (e.g.， Rooker et al.， 2008b).Natural chemical markers， particularly stable oxygenand carbon isotopes (818O and 8C)， have been usedpreviously to determine the natal origin and migrationpatterns of both tropical and temperate tunas (Rookeret al.， 2003， 2008a， 2014； Wells et al.，2012)， and thecurrent application relied on the development of base-line chemical signatures in the otoliths of youngof-the-year (YOY) T. obesus and T. albacares fromdifferent regions in theWCPO.AAftertthe characterization of baseline signatures， region-specificestimates of natal origin were determined by compar-ing 818O and 8’c values in the otolith cores of age-1to age-2+ T. obesus and T. albacares (corresponding tootolith material deposited during the YOY period) toour baseline sample for each species. METHODS YOY T. obesus and T. albacares were collected fromfour geographic regions in the WCPO： (i) far westequatorial (Philippines， Indonesia)， (ii) west equato-rial (Marshall Islands， Solomon Islands)， (iii) centralequatorial (Line Islands， French Polynesia) and (iv)Hawaii [Hawaii (Island)， Maui， Oahu， Kauai， CrossSeamount] (Fig.1). Our baselines were comprised ofYOY T. obesus and T. albacares sampled in late 2007to mid-2008 and late 2008 to mid-2009， referred tohereafter as 2008 and 2009， respectively (Table 1).Within each presumed spawning/nursery area， speci-mens were sampled from multiple collection dates andlocations to obtain a representative baseline samplefor each species. Age-1 to age-2+T. obesus and T. al-bacares of an unknown natal origin were collected in2009-2010 from the west equatorial Pacific (MarshallIslands) and Hawaii regions (Table 1). In the presentstudy， T. obesus and T. albacares ranging in size from20 to 60 cm fork length (FL) and 70 to 120 cm FLwere classified as YOY (~3-12 months old) and age-1to age-2+， respectively (Lehodey and Leroy， 1999；Lehodey et al.，1999). Sagittal otoliths were extracted from both fresh andfrozen T. obesus and T. albacares， cleaned of biologicaltissue and stored dry. One sagittal otolith from eachtuna was embedded in Struers EpoFix resin (StruersA/S， Ballerup， Denmark) and sectioned using a low-speed ISOMET saw (Beuhler， Lake Bluff， IL， USA) to Figure 1. Map of the four regions sampled for bigeye tuna (Thunnus obesus) andyellowfin tuna (T.albacares) in the west-ern and central Pacific Ocean (WCPO)：(a)· farr westequatorial(Philippines，Indonesia)， (b) west equatorial (MarshallIslands， Solomon Islands)， (c) centralequatorial (Line Islands， French Polyne-sia) and (d) Hawaii (Hawaii， Maui，Oahu， Kauai， Cross Seamount). Collec-tion areas for T. obesus (red) and T. al-bacares (black) in the WCPO denoted. Table 1. Summary data for bigeye tuna (Thunnus obesus) and yellowfin tuna (T. albacares) collected in the four regions of thewestern and central Pacific Ocean (WCPO). Region Location(s) Year Age N FL cm (SD) T.obesus A Philippines 2008 YOY 18 27.7(4.1) 2009 YOY 22 30.0(5.4) B Marshall & Solomon Islands 2008 YOY 34 32.8(3.1) 2009 YOY C Line Islands/French Polynesia 2008 YOY 25 48.3(6.2) 2009 YOY 41 45.8(4.4) D Maui/Oahu/Kauai/Cross 2008 YOY 42 51.6(6.1) 2009 YOY 7 54.4(5.1) B Marshall Islands 2009-2010 1-2+ 50 110.7(6.5) D Maui/Oahu/Kauai/Cross 2009-2010 1-2+ 42 75.7(10.8) T. albacares A Philippines/Indonesia 2008 YOY 25 26.0(3.5) 2009 YOY 29 32.2(4.4) B Marshall & Solomon Islands 2008 YOY 50 32.9(3.1) 2009 YOY 16 53.9(2.9) C Line Islands/French Polynesia 2008 YOY 25 52.3(2.2) 2009 YOY 18 46.9(4.1) D Maui/Oahu/Kauai/Cross 2008 YOY 62 44.5(9.2) 2009 YOY 43 39.7(8.0) B Marshall Islands 2009-2010 1-2+ 50 112.9(4.6) D Maui/Oahu/Kauai/Cross 2009-2010 1-2+ 62 79.1(9.4) (A) far west equatorial (Philippines， Indonesia)，(B) west equatorial (Marshall Islands， Solomon Islands)， (C) central equatorial(Line Islands， French Polynesia) and (D) Hawaii (Hawaii， Maui， Oahu， Kauai， Cross Seamount). Collection data for age-1 to 2+T. obesus and T. albacares collected in two regions (west equatorial and Hawaii) are also shown. The mean fork length (cm) andrange provided by region and year. obtain a 1.5 mm transverse section that included theotolith core. Otolith sections were then attached to asample plate using CrystalbondM thermoplastic glue(SPI Supplies/Structure Probe Inc.， West Chester， PA，USA)， and the region corresponding to the early por-tion of the YOY period was isolated and powderedusing a New Wave Research MicroMill (Fremont，CA，USA). Similar to Wells et al. (2012)， the tem-plate was constructed using otolith measurements fromthe smallest individuals [<_25 cm fork length (FL)] inour sample of each species， which represented corematerial accreted during the first ~3 months of life.The final milling template was the same for T. obesusand T. albacares， and a series of drill passes was runover the preprogrammed milling template using a 500-um-diameter Brasseler carbide bit (Brasseler USA，Medical L.L.C.，Ventura， CA， USA) until a depth ofapproximately 800 um was reached. The powderedcore material was transferred to silver capsules forstable isotope analysis. Otolith 8180 and 813c values for YOY T. obesusand T. albacares were determined using an automated carbonate rpreparation deviiccee (KIEL-III；ThermoFisher Scientific， Inc.， Waltham， MA， USA) coupledto an isotope ratio monitoring mass spectrometer (Fin-nigan MAT 252； Thermo Fisher Scientific， Inc.) atthe University of Arizona Environmental Isotope Lab-oratory. Powdered otolith samples were reacted withdehydrated phosphoric acid under vacuum at 70 ℃.The isotope ratio measurement was calibrated basedon repeated measurements of the National Bureau ofStandards (NBS)， NBS-19 and NBS-18， with 6 stan-dards run for every 40 samples； precision was approximately ±0.11%。 and ±0.08%。 for 8o and 8c，respectively. Otolith 818o and 8l’c values arereported relative to the Vienna Pee Dee Belemnite(VPDB) scale after comparison to an in-house labora-tory standard calibrated to VPDB. Elemental chemistry was determined using either asolution-based or laser ablation (LA) approach on aninductively coupled plasma-mass spectrometer (ICP-MS). Because the goal for this component of theinvestigation was to determine whether the additionof elements improved our ability to distinguish YOY T. obesus and T. albacares from different regions， theother sagittal otolith from a subset of the individualsused for otolith 818 and 8c analysis was includedin ICP-MS runs. Solution-based ICP-MS was per-formed on T. obesus otoliths using the Agilent HP4500 quadrupole ICP-MS at the University of Maryland Center for Environmental Science， ChesapeakeBiological Laboratory. The whole otolith approachprovided an integrated measure of the entire YOY per-iod. Otoliths were first cleaned with doubly deionizedwater (DDIH2O) and hydrogen peroxide (H2O2)，immersed briefly (1 min) in dilute nitric acid (HNO3)to remove surface contamination， and then driedunder a laminar flow clean fume hood. In preparationfor instrument analysis， each otolith was weighed tothe nearest 0.01 mg and placed in a clean， plastic cen-trifuge tube. Whole otoliths were then digested in con-centrated HNO， and the quantities of acid used andvolumes were proportional to the otolith weights.Digests were diluted with DDIH2O to a final acid con-centration of 1% HNO3. The levels of Ca and Sr weredetermined using otolith-certified reference material(CRM； Yoshinaga et al.， 2000)， and four elementalratios (Mg：Ca， Mn：Ca， Sr：Ca and Ba：Ca) in the wholeotoliths of YOY T. obesus were estimated according tothe method of standard additions. Instrument preci-sion was assessed by running the CRM every 5 to 10samples， and external precision (relative standarddeviation) for this reference material was as follows：Mg：Ca=10.6%，Mn：Ca=8.7%， Sr：Ca=1.8% andBa：Ca=18.9%. In contrast， otoliths of YOY T. al-bacares were processed with a LA ICP-MS at theWoods Hole Oceanographic Institution Plasma MassSpectrometry Facility. The system consisted of a NewWave Research NWR 213 nm Nd：YAG laser systemand a Thermo Element2 single collector ICP-MS.Ablation diameters were 60 um， and the location ofthe early YOY period was based on otolith microstruc-ture analysis. The first ablation spot was set at the oto-lith core (operationally defined as the narrowest partof the otolith section)， followed by two equally spacedspots on each side of the core (n =3 ablation spots perotolith). Ablation spots incorporated otolith materialaccreted during the first 3-4 months of life. Theablated material was transported via a Helium (He)gas stream to the dual-inlet quartz spray chamberwhere it was mixed with a2% HNO3 aerosol from aself-aspirating PFA20 uL min-nebulizer， usingArgon (Ar) as the sample gas， in the concentric partof the quartz dual-inlet spray chamber. A single isotope was measured from each of the five elements ("Li，25Mg， 48Ca，5Mn， 88Sr and 138Ba). We ran an instrument blank (2% HNO3) and two otolith CRMs every nine samples. We blank-corrected raw isotope valuesby calculating a blank value for each sample using alinear interpolation between bracketed blanks. A dis-solved otolith CRM (Sturgeon et al.， 2005)， diluted toa Ca concentration of 40 ug g二1， was used to correctfor instrument isotope mass bias according to Waltheret al. (2008). Instrument precision was assessed by running another CRM (Yoshinaga et al.， 2000)， similarlydissolved and diluted toa Ca concentration of40 ug g. External precision (relative standard devia-tion) for this reference material (n=19) was as fol-lows： Li：Ca=5.8%， Mg：Ca=1.9%，Mn：Ca =20.1%，Sr：Ca=1.6% and Ba：Ca=2.5%. For both solution-based and laser ablation ICP-MS runs， appropriateprocedural blanks and reference material (e.g.，NIST614， NIST 915a) were used to estimate the recovery，precision and accuracy of analytical runs. Analysis of variance (ANOVA) and multivariateanalysis of variance (MANOVA) were used to test forregional differences in stable isotopes and element：Caratios of YOY T. obesus and T. albacares. MANOVAsignificance was based on Pillai's trace statistic， andsignificance for all ANOVA and MANOVA tests wasbased on an alpha level of 0.05. Canonical discriminant analysis (CDA) was used to display multivariatemeans of otolith chemistry data in a reduced space(two dimensions) for MANOVAs showing a significant region effect. Discriminant function coefficientswere included on CDA plots as biplot vectors from agrand mean to show the discriminatory influence ofstable isotopes or trace elements used in each model.Quadratic discriminant function analysis (QDFA) wasthen used to determine the classification accuracy ofYOY T. obesus and T. albacares to eachnursery(n =4) using otolith chemistry.In addition， QDFAwas used to determine the classification accuracy ofboth species to broader regions (e.g.， west equatorialPacific versus central Pacific). QDFA is the preferredapproach when the variance-covariance matrix of ourpredictor variables differs. In addition， we opted forQDFA because it does not have the homogeneity ofcovariance matrices assumption and is robust to mod-erate deviations from normality (McGarigal et al.，2000). Region-specific estimates of natal origin for T. obe-sus and T. albacares were estimated by comparing oto-lith core 88o and 8c values of age-1 to age-2+tunas to 8180 and 83C values from core material ofYOY tunas in our baseline samples. The natal origin ofT. obesus and T. albacares from two different regionsof the WCPO (the Marshall Islands and the HawaiianIslands) was determined using direct maximum likeli-hoodestimation(MLE)4aanda classification-based estimation (maximum classification likelihood，MCL)from the mixed-stock analysis program HISEA (Mil-lar， 1990). We focus on MLE because the performanceis typically superior to classification-based methods；however， classification-based methods such as MCLappear to be more robust than direct MLE to anoma-lies in baseline data (Millar， 1987，1990)， and thusresults from this estimator are included for compara-tive purposes. Mix-stock analysis was run under thebootstrap mode to obtain standard deviations aroundestimated proportions (error terms) with 500 simula-tions. Prior to mixed-stock analysis， otolith 818o and8C values of YOY (baseline) and age-1 to age-2+T. obesus and T. albacares were plotted in ordinationspace to further evaluate whether all potential sourcepopulations ((i.e.， nurseries)were sampled in theWCPO (Chittaro et al.， 2009). A small percentage ofotolith 8l8 and 8’c values from our sample of age-1to age-2+ tuna occurred outside the 95% confidenceellipse of the YOY baselines for each species (< 2%)，and thus potential bias as a result of the presence ofindividuals from nurseries not sampled was assumed tobe insignificant. RESULTS Variation in YOY signatures Otolith 80 and 8C values for two cohorts (2008and 2009) of YOY T. obesus and T. albacares were dis-tinct among regions in the WCPO. A temporal effectwas observed for otolith 818o， with mean values beingstatistically different between years in one region for T. obesus (far west equatorial) and two regions forT. albacares ((far west equatorial， west equatorial)(ANOVA， P <0.05； Table 2). Otolith 8180 values ofboth species from far west equatorial waters weredepleted by more than 0.2%。 in 2009 relative to 2008.No interannual effect in otolith 8180 was observed foreither species in the central equatorial or Hawaiiregion， with differences between years typically lessthan 0.1% A significant year effect on otolith 8lcwas also detected for T. obesus in the far west equato-rial region， with more enriched values observed in2009 (Table 2). No temporal effect on otolith 8l’cwas observed for T. albacares and all four regions werestatistically similar between 2008 and 2009. Regional differences in otolith 818o for both YOYT. obesus and T. albacares were maintained when2008 and 2009 cohorts were pooled， indicating thattemporal variability was minor relative to geographicvariability. The mean otolith 8o (± 1 SD) forT. obesus varied significantly among the four regions(ANOVA P<0.01) and increased from west to eastin equatorial waters： far west equatorial(-2.65±0.20%)， west equatorial (-2.22±0.22%)and central equatoriall (-1.85±0.19%)(Fig. 2).The Hawaii region (-1.94±0.20%) was most simi-lar to the central equatorial， with the mean differencebeing less than 0.1% and overlapping 95% confidencelimit (CL) ellipses around multivariate means fromCDA (Fig.3). The otolith 81o for YOY T. albacaresalso varied regionally(ANOVA P<0.01) and dis-played a similar pattern to T. obesus with valuesincreasing from west to east： farr west equatorial Table 2. Mean otolith 818o and 813cvalues foryoung-of-theyear((YOY)bigeye tuna(Thunnus obesus) and yellowfin tuna (T. albacares) collected in2008 and 2009 from four regions of thewesternandcentral Pacific Ocean(WCPO). 2008 2009 Difference Region 813c 818o N 8c 818o N sc P 180 P T. obesus -10.29 -2.49 -9.58 -2.76 0.71 ** 0.27 ** -9.77 -2.29 -9.72 -1.88 -9.65 -1.84 0.07 0.04 ns -9.39 -1.96 -10.08 -1.94 0.69 0.02 ns T. albacares -10.56 -2.65 -10.30 _-2.87 0.26 0.22 ** A -9.47 -2.51 -9.32 -2.37 0.15 0.14 ** -9.72 -2.13 -9.61 -2.18 0.11 0.05 ns -9.71 -1.95 -9.74 -1.69 0.03 0.26 ns A = far west equatorial， B= west equatorial，C= central equatorial，D =Hawaii. Mean difference in otolith 8180 and 8c between years is shownalong with results of ANOVA (NS= non significant， **p<0.01). Samplesizes (N) provided. Figure 2. Contour plots of otolith 810 and8C for young-of-the-year (YOY) bigeye tuna (Thunnus obesus) and yellowfin tuna(T. albacares) from the four regions of the western and central Pacific Ocean (WCPO)： (a) far west equatorial， (b) west equato-rial， (c) central equatorial and (d) Hawaii. Bivariate kernel density estimated at four levels (25%，50%，75% and 100%). 1Density (-2.77 ±0.26%)， west equatorial (-2.48±0.21%)and central equatorial (-2.14±0.25). In contrastto T. obesus， otolith 818o for T. albacares in the twocentral regions was different， with the Hawaii region(-1.83±0.26%) being enriched by approximately0.3% relative to the central equatorial region. Multi-variate means and 95% CL ellipses for all four regionswere clearly separated on the CDA plot， and similar toT. obesus， biplot vectors on CDA indicated that regio-nal discrimination was strongly influenced by 8o(Fig.3). Otolith 8c for YOY T. obesus was statisticallysimilar among the four regions (ANOVA P >0.05)，and CDA vectors indicated that 8C was marginallyimportant in separating regions (Fig.3). Mean values(±1SD) for all regions were remarkably similar andwithin 0.2%： far west equatorial (-9.91±0.49%)，west equatorial (-9.78±0.38%)， central equatorial(-9.77±0.49%) aand Hawaii (-9.93±0.43%).While otolith 8lC was significantly different amongthe four regions sampled for T. albap>cares (ANOVAP>0.01)， the effect was due almost entirely to oneregion， the far west equatorial， with CDA plots show-ing that the multivariate mean for this region wasstrongly influenced by 8l’c (Fig. 3). Mean otolithsc for T. albacaresin this region(-10.42±0.59%) was depleted by more than 0.7%/00relative to all other regions sampled： west equatorial (-9.43±0.47%)，central equatorial (-9.67土0.44%) and Hawaii (-9.67±0.43%). QDFA parameterized with otolith 8180 and 8cvalues indicated that the overall cross-validated classi-fication success of YOY T. obesus to the four regionswas 58% in 2008 (range： 48-84% by region). Sampleswere not available for one region (west equatorial) in2009， and overall classification success to three regionsin this year was 78% (range： 66-100%)； QDFA usingpooled 2008 and 2009 data resulted in an overall clas-sification success of 61% (range： 52-80%). Regionaldiscrimination of YOY T. albacares using otolith 8180and 8c was higher with an overall cross-validatedsuccess of 72% (range： 59-86%) and 84% (range： 81-90%) in 2008 and 2009， respectively； QDFA usingpooled 2008 and 2009 data resulted in an overall clas-sification success of 74% (range： 65-80%). For bothspecies， the highest classification success was attribu-ted to either the far west or west equatorial regions.When the western (far west and west equatorial) andcentral (central equatorial， Hawaii) locations werepooled， the overall classification success to these twolarger regions improved to 82% in 2008 (range： 79-85%) and 100% in 2009 for T. obesus and 91% in2008 (range： 88-95%) and 94% in 2009 (range： 89-98%) for T. albacares. Our ability to discriminate YOY T. obesus and T. albacares to regional nurseries wasdue primarily to otolith 8180 (Figs 2 and 3)， and Figure 3. Canonical discriminant analy-sis (CDA) based on otolith 8o and8l’c for young-of the-year (YOY) bigeyetuna (Thunnus obesus) and yellowfin tuna(T. albacares) from the four regions of thewestern andcentral PacificOcean(WCPO)： (A) far west equatorial， (B)west equatorial， (C) central equatorialand (D) Hawaii. Ellipses represent 95%confidence limit around each multivari-ate mean. Biplot vectors from a grandmean indicate the influence of 8180 and813con regional discrimination. T. albacares discrimination using this marker alone resulted ina minor reduction to overall classification successto the four regions： T. obesus((2%) andT. al-bacares (9%). Spatial variation in otolith element：Ca ratios ofYOY T. obesus and T. albacares collected in 2008 wasalso present among the regional nurseries (MAN-OVA， P <0.01， Table 3 and Fig. 3). The Mg：Ca andBa：Ca ratios in the otoliths of T. obesus from the farwest equatorial region were significantly higher thanobserved for individuals from all other regions sampled(Tukey’s HSD， P <0.05)， and CDA vectors supportthe importance of both as being influential in discrimi-nating T. obesus from this region(Fig. 4). Meanvaluesof Mg：Ca (639 umol mol-)andBa：Ca(1.0 umol mol) from the far west were at least twofold higher than values observed in the three otherregions： Mg：Ca (range： 199-310 umol mol-) and Ba：Ca range (range： 0.4-0.5 pmol mol-). Similarly， oto-lith Mn：Ca was also highest for T. obesus from far westequatorial waters (3.1 umol mol-)； however， otolithMn：Ca was statistically similar for individuals from theother equatorial regions but different from individualsfrom the Hawaii region (1.9 umol mol-l) (Tukey’sHSD， P<0.05). Elemental signatures in the otolithcores of T. albacares also varied among the four regions(MANOVA， P <0.01)， with multivariate means and95% CL ellipses from CDA most similar for the farwest equatorial and west equatorial regions. Only twoof the five element：Ca ratios examined were deter-mined to be influential (Li：Ca and Sr：Ca). Meanotolith Li：Ca for T. albacares was significantly lower in Region Mg：Ca Mn：Ca Sr：Ca Ba：Ca N T. obesus A 638(158) 3.1(1.5) 1752(104) 0.960(.54) 10 B 310(250) 2.1(0.7) 1731(210) 0.42(0.21) 25 C 199(62) 2.7(1.8) 1698(200) 0.51(0.49) 16 D 259(197) 2.0(0.6) 1771(176) 0.41 (0.55) 34 Region Li：Ca Mg：Ca Mn：Ca Sr：Ca Ba：Ca N T. albacares A 5.6(0.8) 266(120) 2.3(0.7) 1750(80) 0.38(0.14) 14 B 5.4(1.0) 208(153) 2.3(1.5) 1786(104) 0.41(0.10) 24 C 4.5(1.1) 125(40) 5.5(1.9) 1627(91) 0.51(0.19) 12 D 5.1(1.3) 121(324) 2.3(1.8) 1620(282) 0.39 (0.14) 31 A= far west equatorial， B = west equatorial， C =central equatorial， D=Hawaii. All ratios expressed as umol mol-and measurements were solutionbased (whole otoliths) for T. obesus and laser ablation ICP-MS(otolith cores)for T. albacares. Li：Ca not determined for YOY T. obesus. Sample sizes (N)provided. ( Table 3 . M ean otolith element：Ca r a tios of young-of-the year (YOY) bigeye tuna(Thunnus obesus) and yellowfin tuna (T . albacares) collected in 2008 f rom fourregions of the e West Central I P acific Ocean (WCPO). ) the central equatorial (4.5 pmol mol-) relative tothe three other regions investigated(range： 5.1-5.6 umol.mol-) (Tukey's HSD， P <0.05). The oto-lith Sr：Ca also varied regionally with mean values forT. albacares from the far west and west equatorialwaters((1750 and 1786 umol mol-， respectively)being greater than values observed for individuals fromthe central equatorial (1627 umol mol-) or Hawaii(1620 pmol mol-) regions. For the other threeelement：Ca ratios examined， mean values were statis-tically similar among the four regions： Mg：Ca (range：125-265 umol mol-)， Mn：Ca (range： 2.3-3.5 umolmol-) and Ba：Ca (0.4-0.5 umol mol-). QDFA parameterized with element：Ca ratios in theotoliths of T. obesus and T. albacares from a singleyear ((2008) indicated that regional discriminationusing these markers was lower relative to models basedon otolith 8180 and 813C. Cross-validated classifica-tion success for models based on 8180 and 8c versuselement：Ca ratios was higher for both T. obesus (60%versus 46%) and T. albacares (68% versus 62%).Regional discrimination improved by adding elemen-tal data to models with 818o and 8lc； nevertheless，overall cross-validated classification success to the fourregions only improved slightly with the addition of theelements； T. obesus (61%) and T. albacares (72%).Given that the addition of elements only improvedregional discrimination by 1% to 4% regardless of theICP-MS approach used [bulk chemistry (whole otolith) versus laser ablation (otolith core)]， 818o appearsto represent the most important marker for distin-guishing YOY T. obesus and T. albacares from regionalnurseries in the WCPO. Natal origin The otolith core 818o and 81’c values of age-1 andage-2+T. obesus and T. albacares collected in 2009from the west equatorial (Marshall Islands) andHawaii regions were used to assess the significance oflocal production and movements for both species.YOY T. obesus and T. albacares collected in 2008were used as the baseline sample for mixed-stock predictions， which allowed for age-class matchinga frac-tion of our sample. Age-1 (c. 70-100 cm FL) T. obesusand T. albacares collected in 2009 represented individ-uals that were ‘age-class matched'to the 2008 baseline，and accounted for approximately half (53% and 50%，respectively) of our sample， with the remaining age-2+tuna probably produced in 2007 or one year earlierthan the baseline. Direct MLE estimates for T. obesus(n=50) and T. albacares (n =50) from the westequatorial waters (Marshall Islands) indicated a lim-ited contribution from other regions (Fig. 5). The pre-dicted contribution(MLE % ±1 SD) of T. obesusfrom the three other regions to the west equatorialregion was nil (0% ±0%) and all recruits originatedfrom nurseries in west equatorial waters， highlightingthe importance of local production. Similarly， nearlyall of the T. albacares in the west equatorial samplewere produced in this region (95% ±6%)， with aminor contribution (5%) derived from the centralequatorial region (Line Islands， French Polynesia).MCL estimates were identical for T. obesus(100%from a local source) but indicated that the contribu-tion from the central equatorial region may be higher(24%±7%) than predicted with MLE. Figure 4. Canonical discriminant analy-sis (CDA) based on otolith trace ele-ments for young-of-the-year (YOY)bigeye tuna(7(Thunnus obesus) and yellowfin tuna (T. albacares) from the fourregions of the western and central PacificOcean (WCPO)： (A) far west equatorial，(B) west equatorial， (C) central equato-rial and (D) Hawaii. Ellipses represent95% confidence limit around each multi-variate mean. Biplot vectors from a grandmean indicate the influence of each ele-ment：Ca ratio on regional discrimination. T.albacares Mn/Ca Mg/Ca Ba/Ca Sr/Ba Li/Ca Local production was also important for age-1 toage-2+ T.albacares (n=62) caught in the Hawaiiregion， and direct MLE estimates indicated that indi-viduals in our Hawaii sample were exclusively pro-duced in the Hawaii region (100%±0%)， with nocontribution from any of the equatorial nurseries(Fig.5). In contrast， the primary source of T. obesus(n =42) in Hawaii was the central equatorial region(65% ±35%)， indicating the potential for significantnorthward movement of young T. obesus from areassouth of Hawaii.Only about a third (34% ±35%) ofthe T. obesus in our Hawaii sample were linked toproduction from this region. No migrants from the farwest or west equatorial regions were detected in theHawaii sample for either species. It should be notedthat error terms from direct MLE estimates of the Hawaii sample were relatively large (c. 35%)， indicating a high degree of uncertainty associated with ourpredictions of the natal origin of T. obesus in thisregion. Again， MCL estimates matched MLE for thespecies showing 100% production from local sources(T. albacares)； however， MCL indicated that the con-tribution from local sources in the Hawaii region maybe slightly higher (42% ±36%) than predicted withMLE. DISCUSSION Temporal variability in otolith8180 and 8c wasdetected for YOY T. obesus and T. albacares collectedin 2008 and 2009， but interannual differences werelimited to the far west and west equatorial regions. Temporal shifts in 8O for biogenic carbonates arepredictably linked to changes in ambient seawater8180 and temperature (Grossman， 2012). A moder-ately strong La Nina event was observed in 2007-2008and this phenomenon resulted in heavy precipitationthroughout the far western equatorial region (Lyonand Camargo， 2009； Gordon et al.， 2011). In fact， dur-ing the winter of 2007-2008， the Philippines experi-enced the greatest rainfall in nearly four decades， withpeak precipitation occurring in early 2008 (Pullenet al.， 2015) or after all individuals in our 2008 samplewere collected from this region. Increased precipita-tion in early 2008 from the La Nina event combinedwith high precipitation again in late 2008 (Pullenet al.， 2011) resulted in lower salinities in coastal areas(Gordon et al.， 2011). In turn， it appears that bothT. obesus and T. albacares in our 2009 sample wereexposed to seawater with depleted 81o values (rainwater is depleted in O relative to seawater)， which isin accord with the depleted otolith 80 values(~0.2%)observed in 2009. Apart from salinity(8o)， a sea surface temperature (SST) anomaly wasalso observed between 2008 and 2009， with meanSSTs during the period approximately 4 months priorto capture (September-December) being about 0.5 ℃warmer in 2009 compared to the same period in 2008(based on cumulative monthly SST data fromMODIS/Aqua Global Satellite Level 3， Daytime SSTaccessed using Marine Geospatial Ecology Toolsextension in ArcGIS 10.2；Roberts et al.，2010). Tem-perature dependence of 810 is well documented inotoliths and other biogenic carbonates (Grossman andKu， 1986)， and our finding of depleted otolith 818owith increasing temperature in the Philippines in 2009is in agreement with the well-documented inverserelationship between otolith 8 and water tempera-ture (Thorrold et al.， 1997； Hoie et al.， 2003). In response， the combined effects of ocean freshening(lower 81O) and warmer SST in 2009 for the far westequatorial region appeared responsible for observedtemporal shifts in otolith 8180 of both T. obesus andT. albacares. Previous studies have reported that the interannualvariability in otolith 81C also occurs for tunas in boththe Atlantic Ocean (Rooker et al.，2008b) and PacificOcean (Wells et al.， 2012)， but differences are oftenvariable and add little to the regional discriminationof individuals using these natural tracers (Rookeret al.，2014). For all four regions， otolith 8l’c of indi-viduals collected in 2008 and 2009 were statisticallysimilar for YOY T. albacares. We did observe a yeareffect for YOY T. obesus from two regions (far westand central equatorial). Nevertheless， as reported pre-viously， linking seawater l’c directly with otolith8C is problematic because dissolved inorganic car-bon (DIC) in seawater and a variety of others factors(diet， metabolism and kinetic effects) can ultimatelydetermine the otolith 83c values (Hoie et al.， 2003；McMahon et al.，2013). Regional variation in otolith 810 was observed forboth YOY T. obesus and T. albacares and maintainedeven when the 2008 and 2009 cohorts were pooled，signifying that regional differences were sufficientlylarge enough to overshadow temporal variabilitywithin or among the regions examined. Similar to pre-vious research on Atlantic bluefin tuna (T. thynnus；Rooker et al.， 2014) and Pacific bluefin tuna (T. orien-talis， Shiao et al.， 2010)， discrimination of YOYT. obesus and T. albacares to different geographicregions in the WCPO was due primarily to one mar-ker， otolith 8180.Otolith 8180 values of both specieswere the most depleted in the far west equatorialregion and11increased1 moving eastward1into cenetral equatorial waters. Similar to the aforementioned explanation of interannual variability， regional differ-encesin seawater 8o and temperature appearresponsible for observed differences in otolith 818o8camong the four regions. Seawater880 values fromthe Global Seawater Oxygen-18 Database (v.I"http：//data.giss.nasa.gov/o18data/) and otolith 8oof T. obesus and T. albacares followed the same pat-tern in the WCPO. In general， otolith 810 values ofboth species increased moving eastward along theequatorial Pacific (lowest to highest： far west， west andcentral equatorial). Similarly， both seawater 818o(LeGrande and Schmidt，2006) and sea surface salinity(Delcroix et al.，2011) followed the same west-to-eastenrichment pattern continuing northward to theHawaii region： far west equatorial (lowest)， west equa-torial， central equatorial and Hawaii (highest). Themean difference in otolith 8O between the two geo-graphic end points of our sampling design， the far westeFawquatorial and Hawaii regions， was approximately~0.8-1.0%。 with otolith 81Oof YOY tunas being mostenriched in the higher salinity waters off the HawaiianIslands. Otolith 8180 is also a known indicator of watertemperature experienced by fishes (Kalish， 1991；Thorrold et al.，1997)， and its value as a paleother-mometer 1Sisaa result of the strong temperaturedependence of 180partitioning between the DIC pooland seawater (Grossman， 2012). Therefore， spatialvariation in SST among the four regions probably con-tributed to the observed patterns in otolith 8180 ofYOY T. obesus and T. albacares. Recently， Kitagawaet al. (2013) established a relationship between ambient water temperature and otolith 8180 for larval Paci-ficbluefintuna(T. orientalis)andt'the observedrelationship was a 0.27%。 decrease in otolith 818owith every 1℃ increase in water temperature overthe range investigated. Moreover， comparable temper-ature-dependent relationships (i.e.， slopes) have beenreported for several other taxa of marine fishes (e.g.，Kalish， 1991； Hoie et al.，2004). The estimated cumu-lative annual mean SST varied markedly among thefour regions in 2008 (far west equatorial 29.9°℃， westequatorial 229.1℃， central equatorial 26.3℃， andHawaii 25.3℃)， with the two regional end members(far west equatorial and Hawaii) differing by approxi-mately 4.0 ℃ (based on cumulative monthly SSTdata at 10 random locations within each samplingregion from MODIS/Aqua Global Satellite Level 3).Based on the temperature-otolith 8180 relationship byKitagawa et al. (2013)， this would correspond to anincrease of approximately 1.0% in the cooler waters ofthe Hawaii region relative to the far west equatorialregion. Observed differencesinotolithhs8o cof T. obesus and T. albacares between the two geographicend points was approximately 0.8% and 1.0%， respec-tively， with the individuals from Hawaii showingenriched otolith 8oas predicted by the otolith818O-water temperature relationship. Seawater 818values from each of the regional collection locationsare needed from 2008 and 2009 to estimate the rela-tive impact of temperature dependence (Kim et al.，2007). Unfortunately， these data are not available，which limits our ability to assess the relative impor-tance of seawater 8180 and temperature on observedotolith 8180 of T. obesus and T. albacares. As noted previously， linking seawater 8 C directlywith otolith 813C is problematic because other factorscan ultimately influence the otolith values (Hoieet al.， 2003). Similar to other studies examining theotolith 8180 and 813c of tunas at the ocean basin scale(e.g.， Wells et al.， 2012； Rooker et al.， 2014)，， weobserved that spatial variation in otolith 8c wasminor to inconsequential， even though this markerhas shown promise for examining the origin and move-ment of juvenile fishes from estuarine or coastal nurs-eries (Thorrold et al.， 2001； Mateo et al.， 2010).Otolith 813c was not influential in discriminatingYOY T. obesus from the four regions of the WCPOand apart from one region (far west equatorial)， otolith81c of YOY T. albacares was similar among the otherthree regions sampled for this species. Discriminationof YOY T. obesus and T.albacares using this markeralone contributed little to the overall classificationsuccess to regional nurseries， suggesting that its valuefor distinguishing pelagic fishes from different watermasses within the WCPO is limited. Elemental ratios in otoliths have also shown poten-tial for assessing the origin and movement of estuarine，coastal and open ocean fishes (e.g.， Gillanders andKingsford， 1996； Arai et al.， 2005；Elsdon and Gillan-ders， 2006). In the present study， both elements andstable isotopes were quantified for a subset of individuals to evaluate the discriminatory power of each foridentifying the origin of YOY tunas inhabiting theWCPO. Spatial variation in certain elements was pre-sent for both YOY T. obesus and T. albacares， withseveral element：Ca ratios being noticeably higher incertain regions. For T. obesus， otolith Mg：Ca， Ba：Ca，Mn：Ca ratios were elevated in the far west equatorialregion， and observed differences were not entirelyunexpected because hydrography and trace elementfluxes vary regionally in the WCPO. Our T. obesussample from the far west equatorial region was basedon tuna collected in the southern Philippines (MoroGulf). This region is heavily influenced by riverineinputs from the Rio Grande de Mindano， which has a drainage area of over 23 000 km’ and contains highloads of certain metals (Breward et al.， 1996)， poten-tially explaining the observed increase in certain ele-ments (Mg，Mn) in the otolith of T. obesus from thisregion. Concentrations of trace metals in the other，more open ocean regions of the equatorial Pacific(west and central regions) were lower and this is prob-ably because metal concentrations from anthropogenicand lithophilic sources decrease precipitously as thedistance from the coastline increases (Bruland andLohan， 2004). Spatial variation was also detected forYOY T. albacares but different element：Ca ratios var-ied regionally (Li：Ca， Sr：Ca)， possibly indicating spatial or temporal partitioning of areas within regionalnurseries by the two congeners. Pronounced regional variation in otolith oresulted in respectable classification success to putativenurseries for both YOY T. obesus and T. albacares inthe WCPO. Models parameterized with otolith 8Oand 81’c indicated that classification to the four regio-nal nurseries was greater for T. albacares than T. obe-sus. The cross-validated success of T. albacares in 2008and 2009 (72%and 84%， respectively) was slightlyimproved over estimates by Wells et al. (2012)， whichdid not follow the geographic groupings used in thecurrent study. The classification success of YOYT. obesus to the four regions was lower because otolith81o values in two of the four regions (central equato-rial， Hawaii) overlapped significantly， resulting in ahigher number of misclassified individuals and greateruncertainty in mixed-stock predictions of natal origin. Although otolith chemistry is widely used to assessthe origin of tunas and other oceanic fishes， most ofthe studies to date have focused on 8180 and 8C signatures (e.g.， Wells et al.，2010； Rooker et al.， 2014；Fraile et al.， 2015) rather than element：Ca ratios，either alone or combined with stable isotopes (e.g.Rooker et al.， 2003). We quantified both stable isotopes and element：Ca ratios for a subset of individuals，and this approach shed light on the resolving power ofthese two classes of natural markers for open oceanspecies. The cross-validated classification success formodels based on stable isotopes only versus element：Ca ratios only were always higher for models parame-terized with 818 and 8C， and this finding was con-sistent between thee two species.While overallclassification successto the differeenntt geographicregions improved by adding elemental data to modelsalreaddyy pParameterizedwithstableiisotopes，increase in the resolving power was negligible (< 5%).As a result， the effort and cost to incorporate elemen-tal ratios into otolith chemistry baselines for bothT. obesus and T. albacares may not be warranted for future assessments of tuna origin and mixing in theWCPO. . By comparing otolith 81 and 8C values of age-1 and age-2+ T. obesus and T. albacares to our YOYbaseline， we determined that both local productionand movements influenced the regional compositionof fisheries in equatorial waters of the WCPO. For ourwest equatorial samples of both T. obesus and T. al-bacares， all or nearly all individuals (based on MLE：100% and 95%， respectively) were predicted to be pro-duced from the same region， highlighting the impor-tance of localproduction and suggestingthatlongitudinal movements may be limited in this region.Using both conventional and archival tags， Schaeferand Fuller (2009) investigated the movements and dis-persion of T. obesus in the equatorial eastern Pacificand observed restricted movements and limited disper-sion， which is in accord with our finding of retentionwithin the west equatorial waters. More recent studiesby Schaefer et al. (2015) in the equatorial centralPacific showed regional fidelity but also extensive east-ward longitudinal dispersion that was not observed inthe present study (i.e.， no contribution from far westequatorial waters). Other tagging studies in the CoralSea (Hampton and Gunn， 1998； Clear et al.， 2005)also reported restricted movements and high residencyfor T. obesus. Similar to T. obesus， tagging studies onT. albacares in equatorial waters also indicated limitedlongitudinal dispersion (Sibert and Hampton， 2003)，supporting our finding that nearly all T. albacares inour west equatorial sample were locally produced. Aminor contribution of T. albacares recruits from cen-tral equatorial waters highlights the potential forextendedlongitudinalimovements(> 1000 nmi)，which is known to occur for this species (e.g.， Davieset al.， 2014). Our estimates of the natal origin for T. obesus col-lected in the Hawaii region indicated the potential forahigh degree of south-to-north movement. ForT. obesus caught in the Hawaii region， otolith chem-istry indicated that the primary source of recruits(65%) was not local but rather south of Hawaii， withotolith 80 and 8C values matching those of YOYtuna collected in the central equatorial region (LineIslands， French Polynesia). This suggests that largenumbers of T. obesus migrate northward from south-ern spawning areas with similar water 880 values(e.g.， north of 10°N)； however， it is important toacknowledge that baseline signatures of individualsfrom Hawaii and the central equatorial regions over-lapped， which increased the uncertainty of our predictions (error term 35%). While the majority of age-1and age-2+T. obesus in our Hawaii sample were sourced to central equatorial waters， the area betweenthese two regions (10°N to 18°N) may represent a pro-duction zone or source of recruits. If individuals fromthis region possess otolith 8O values similar to YOYT. obesus from central equatorial waters， predictionsof natal origin to this region will be overstated. Thepotential for an ecologically meaningful contributionof recruits from equatorial nurseries to the Hawaiiregion has been suggested previously for tropical tunas(Hampton and Fournier， 2001) because productionand catch rates in equatorial waters are considerablyhigher than all other regions in the WCPO (Schaeferet al.， 2015). However， a large-scale tagging study bySchaefer et al. (2015) in equatorial waters (releasesfrom 5°S to 8°N) showed constrained latitudinal dis-persion of T. obesus within equatorial waters， withonly a few individuals migrating north of 10°N andinto the Hawaii region. While the northward move-ment from the production zone(s) in the south appearsto explain our otolith chemistry results， it is difficult todetermine the specific latitudinal origin (e.g.， south ornorth of 10°N) of migrants in our Hawaii sample giventhe spatial structure of our baseline sample. We alsoobserved that approximately one-third of the T. obe-sus in our sample were predicted to originate from theHawaii region or rather from local production butagain our error term is relatively high. Spawning ofT. obesus occurs over a widespread region of theWCPO， and ichthyoplankton surveys conducted inthe Hawaiian Islands have collected T. obesus larvae，albeit numbers of larvae are considerably lower thanT. albacares (Lobel and Robinson， 1988； Nikaidoet al.， 1991； Paine et al.， 2008). The presence ofT. obesus larvae around the Hawaiian Islands clearlydemonstrates that spawning can extend above 15°Nand into Hawaiian waters where SST seasonally riseabove the spawning threshold (23-24℃) for this spe-cies (Schaefer et al.， 2005). However， spawning ofT. obesus in this region appears to be centered southof the Hawaiian Islands， midway between Hawaii andthe Line Islands (Nikaido et al.， 1991； Schaefer et al.，2005). While localized spawning or nursery areas inthe Hawaiian Islands or possibly from waters immedi-ately to the south the archipelago (south of 19N)appears to contribute a significant proportion of theT. obesus to this region， production farther southappears to be the principal contributor of recruits tothe Hawaii region. In contrast to T. obesus， Hawaii was identified asthe only source of age-1 to age-2+ T. albacares in oursample from this region， suggesting that Hawaiianwaters serve as both a production zone and/or nurseryarea for T. albacares. Our results are in accord with tagging studies on T. albacares that report restrictedmovements of this species in the Hawaiian Islands(Dagorn et al.， 2007). In fact， Itano and Holland(2000) reported that mean displacement distances ofyoung T. albacares in Hawaii were typically less than50 km from the release location.High fidelity and lim-ited movement of T. albacares in Hawaiian watersmay be due to the quality/quantity of prey resourcesand highly suitable habitats (e.g.， banks， seamounts)in the region (Wells et al.， 2012). Prey availability isoften a primary determinant of the movement for pela-gic fishes and other marine vertebrates (Forcada et al.，2009； Block et al.， 2011； Golet et al.， 2013)， and thehighly suitable habitat experienced by T. albacares inHawaiian waters probably enhances residency and lim-its their movement. Also， T. albacares present in theHawaiian Islands display a lengthy spawning seasonfrom spring to fall when reproductively active adultsand larvae have been documented， possibly eliminat-ing the need to migrate elsewhere (Itano， 2000； Paine，et al.，2008).Regardless of the mechanism responsiblefor retention， our results show that T. albacares col-lected in the Hawaii region were from local sources，indicating that the Hawaii-based fishery for this spe-cies is potentially supported by local production withlittle or no subsidies from equatorial production zonesor nurseries. ACKNOWLEDGEMENTS Funding for this work was provided by the Universityof上Hawai'iPelagic Fisheries Research1Program(JIMAR project 651106 to JRR). A special thanks toG. Castrence， P. Conley， J. Dettling， D. Dettman， B.Fukuda， D. Fuller， E-J Kim， K. Lind， J. Muir， B. Muller，K. Pollock， K. Schaefer， D. Secor， S. Tobiason， R.Wingate and T. Usu. REFERENCES Arai， T.， Kotake， A.， Kayama， S.， Ogura， M. and Watanabe， Y.(2005) Movements and life history patterns of the skipjacktuna Katsuwonus pelamis in the western Pacific， as revealedby otolith Sr： Ca ratios. J. Mar. Biol. Assoc. U.K. 85：1211-1216. Block， B.A.， Jonsen， I.D.， Jorgensen， S.J. et al. (2011) Trackingapex marine predator movements in a dynamic ocean.Nature 475：86-90. ( Breward， N. ， Apostol， A. ， Appleton， J.D.， Gomez， R . a nd Miguel， J. (1996) M ercury and d o ther 上 heavy-metal contamination a s sociated w ith gold mining in t h e A g usan river catchment， Mindanao， the Philippines. BGS OverseasGeology Ser. Tech. Rep.WC/96/61/R：40.pp. ) Bruland， K.W.and Lohan， M.C. (2004) Controls of trace metalsin seawater. In： Treatise on Geochemistry - Volume 6： The Oceans and Marine Geochemistry. H. Elderfield (ed)， NewYork： Elsevier Inc， pp. 23-47. Chittaro， P.M.， Finley， R.J.and Levin， P.S. (2009) Spatial andtemporal patterns in the contribution of fish from theirnursery habitats. Oecologia 160：49-61. Clear， N.P.， Evans， K.， Gunn， J.S.， Bestley， S.， Hartmann，K. and Patterson， T. (2005) Movement of bigeye tuna(Thunnus obesus) determined from archival tag light-levels and sea surface temperatures. In： Migration andPreferences， Habitat of Tuna， Bigeye， Thunnus obesus，on the Coast， East of Australia. Report FRPC， CSIRO1999/109， pp.19-46. Dagorn， L.， Holland， K.N. and Itano， D.G. (2007) Behavior ofyellowfin (Thunnus albacares) and bigeye (T. obesus) tuna ina network of fish aggregating devices (FADs). Mar. Biol.151：595-606. Davies， N.， Harley， S.， Hampton， J. and McKechnie， S. (2014)Stock assessment of yellowfin tuna in the western andcentral Pacific Ocean. Western and Central Pacific FisheriesCommission， Document number WCPFC-SC10-2014/SA-WP -04. 112 pp. Delcroix， T.， Alory， G.， Cravatte， S.， Correge， T. andMcPhaden， M.J. (2011) A gridded sea surface salinity dataset for the tropical Pacific with sample applications (1950-2008). Deep-Sea Res. Pt. 1.58：38-48. Elsdon， T.S. and Gillanders， B.M. (2006) Identifying migratorycontingents of fish by combining otolith Sr： Ca withtemporal collections of ambient Sr： Ca concentrations. J.Fish Biol. 69：643-657. Ferriss， B.E. and Essington， T.E. (2011) Regional patterns inmercury and selenium concentrations of yellowfin tuna(Thunnus albacares) and bigeye tuna (Thunnus obesus) in thePacific Ocean. Can. J. Fish Aquat. Sci. 68：2046-2056. Forcada， J.， Malone， D.， Royle， J.A. and Staniland， I.J. (2009)Modelling predation by transient leopard seals for anecosystem-based management of Southern Ocean fisheries.Ecol. Model. 220：1513-1521. ( Fraile， I. ， Arrizabalaga ， H. and Rooker， J.R. (2015)O r igin o f Atlantic bluefin tuna (Thunnus t hynnus) in the B ay o f Biscay. ICES J. Mar. Sci. 72：625- 6 34. ) ( Gillanders， B.M. and K i ngsford， M.J. (1996) Elements in otolith s may y e lucidate the c ontribution of e stuarine recruitment t o s ustaining c oastal r eef p opulations o f a temperate reef fish. Mar. Ecol. Prog. Ser . 141：13-20. ) Golet， W.J.， Galuardi， B.， Cooper， A.B. and Lutcavage， M.E.(2013) Changes in the distribution of Atlantic Bluefin Tuna(Thunnus thynnus) in the Gulf of Maine 1979-2005. PLoSONE 8：e75480. Gordon， A.L.， Sprintall， J. and Ffield， A. (2011) Regionaloceanography of the Philippine archipelago. Oceanography24：14-27. Grewe， P.M. and Hampton， J. (1998) An assessment of bigeye(Thunnus obesus) population structure in the Pacific Ocean，based on mitochondrial DNA and DNA microsatelliteanalysis. CSIRO Mar. Res. No. 34：29. Grewe， P.M.， Feutry， P.， Hill， P.L. et al. (2015) Evidence ofdiscrete yellowfin tuna (Thunnus albacares) populationsdemands rethink of management for this globally importantresource. Sci. Rep.5：16916. Grossman， E.L. (2012) Oxygen isotope stratigraphy. In： TheGeological Time Scale 2012. F. Gradstein， J. Ogg， M.Schmitz & G. Ogg (eds)Oxford， UK： Elsevier，pp. 181-206. ( Grossman， E . L. and Ku， T.L. (1986) Oxygen an d carbon isotope f ractionation in biogenic a ragonite： t emperature effects.Chem. Geol.. Isotope Geoscience Section 59：59-74. ) ( Hampton， J . and F ournier ， D.A. ( 2001) A s patiallydisaggregate， d， l ength-based， age-structured d p opulation model o f yellowfin tuna ( T hunnus albacares) in th e western and central Pacific Ocean. M ar. Freshwater Res. 5 2：937-963. ) ( Hampton， J . a nd G unn， J. ( 1998) Exploitation and movementsof yellowfin tuna ( Thunnus a l bacares) and bigeye tuna (T . obesus) tagged i n t he n orth-western C oral Sea. M a r. F reshwater Res. 49：475-489. ) ( Harley， S.， Davies， N .， Hampton， J. and Mckechnie，S . (2014) S tock a ssessment of bigeye tuna i n the western a n d centralPacifi c Ocean. In： WCPFC S C10- SA-WP-01， Majuro， Republic of the Marshall Islands， 6-14 August 2014. ) ( H oie，H . ， Folkvord， A. a n d Otterlei， E . ( 2003) E ffect of somaticand o tolith g rowth rate o n s t able isotope c omposition o f early juvenile cod (Gadus morhua L.) otoliths. J. E x p. M a r. Biol. Ecol. 289：41- 5 8. ) ( Hoie， H . ， Otterlei， E. and Folkvord， A. (2004) T emperature-dependent fractionation of stable oxygen isotopes in otolithsof juvenile cod (Gadus morhua L .). ICES J . Mar. Sci. 61：243-251. ) ( Itano， D.G. (2000) T he reproductive biology o f yellowfin tuna (Thunnus a lbacares) in Hawaiian waters and the western t ropical Pacific Ocean： p roject summary. In ： S O EST. P ubl00-01， Joint Institut e for Marine and Atmospheric R esearch ( JIMAR) Contribution. Honolulu，HI： JIMAR，pp. 00- 3 28. ) ( Itano， D.G . and H olland， K .N. (2000) M ovement a n dvulnerability of bigeye (Thunnus o besus) and y ellowfin t una (Thunnus albacares ) in relatio n to FADs a nd n a tural aggregation points. Aquat. L iving Resour. 13：213-223. ) ( J oseph， J .， Squires ， D. ， Baylif f ， W. and Groves， T . ( 2010) Addressing t he problem o f e x cess fishing capacity in tuna f isheries.In： Conservationand M anagement of Transnational Tuna F isheries.R. Allen， J . Joseph & D . S quires ( eds)： Blackwell P ublishing， pp. 1 1 -38. ) Kalish， J.M. (1991) Oxygen and carbon stable isotopes in theotoliths of wild and laboratory-reared Australian salmon(Arripis trutta). Mar. Biol. 110：37-47. Kim， S.T.， ONeil， J.R.， Hillaire-Marcel， C. and Mucci， A.(2007) Oxygen isotope fractionation between syntheticaragonite and water： influence of temperature and Mg 2+concentration.Geochim. Cosmochim. Ac. 71：4704-4715. Kitagawa， T.， Ishimura， T.， Uozato， R. et al. (2013) Otolith 818O of Pacific bluefin tuna Thunnus orientalis as an indicator ofambient water temperature. Mar. Ecol. Prog. Ser. 481：199-209. LeGrande， A.N. and Schmidt， G.A. (2006) Global gridded dataset of the oxygen isotopic composition in seawater. Geophys.Res. Lett. 33：L12604. Lehodey， P. and Leroy， B. (1999) Age and growth of yellowfintuna (Thunnus albacares) from the western and centralPacific Ocean as indicated by daily growth increments andtagging data. WP YFT-2. SCTB 12：16-23. Lehodey， P.， Hampton， J. and Leroy， B. (1999) Preliminaryresults on age and growth of bigeye tuna (Thunnus obesus)from the western and central Pacifc Ocean as indicated bydaily growth increments and tagging data. Sec. Pac. Com.，Oceanic Fish. Prog.， 12th meeting， Stand. Com. TunaBillfish，BET-2：10 pp. ( Lobel， P.S. and Robinson， A.R. ( 1988) Larval f ishes and zooplankton i n a c yclonic eddy i n H awaiian waters. J. P lankton Res. 10：1209-1223. ) ( Lyon， B. and Camargo， S . J. ( 2 009) The se a sonally-varyinginfluence of ENSO on ra i nfall and t r opical cyclone activityin the Philippines. Clim. Dynam.32：125-141. ) ( Mateo， I.， Durbin， E.G.， Appeldoorn， R.S. et al (2010) R ole of mangroves a n urseries forFrench g runt Haemulon flavolineatum and s choolmaster Lutjanus apodus a s sessed b y otolith elemental f ingerprints. Mar. E Ecol. I Prog.Ser. 402：197-212. ) ( McGarigal ， K. ， Cushman， S. andi d S Stafford， S ( . 2000) Multivariate S tatistics for Wildlife and E cology R e search. New York， NY： Springer-Verlag. ) ( McMahon， K.W.， Ha m ady， L.L . and Tho r rold， S. R . (2 0 13) A review of e c ogeochemistry approaches to estimating movements of marine a n imals. Limnol. O ceanogr. 5 8：697-714. ) ( M illar， R.B. (1987) Maximum likelihood es t imation of mi x edstock f ishery composition. Can. J . Fish Aquat. S c i. 44：583-590. ) ( Millar， R.B. (1990) Comparison of methods f o r e s timatingmixed s tock fishery composition. C an. J. F ish A quat. S ci. 47：2235-2241. ) ( Nikaido， H.， Miyabe， N. and Ueyanagi， S. (1991) Spawningtime a nd frequency of bigeye tuna， Thunnus obesus. Nat.Re s . Inst. Far Seas Fish. Bul l .28：47-73. ) ( Paine， M.A.， M cDowell， J.R. a n d G r aves， J.E. (2008) S p ecificidentificatio n usin g COI s equence a n alysis o f s c ombridlarvae collected off the Kona coast of Hawaii Island. Ichthyo. Res. 55：7-16. ) ( Pullen， J.D.， Gordon， A.L.， Sp rintall， J. ( e t al. (2011) Atmospheric a n d o c eanic p r ocesses in the v i cinity of a nisland strait. Oceanography 24 ： 112-121. ) Pullen， J.， Gordon， A.L.， Flatau， M.， Doyle， J.D.， Villanoy， C.and Cabrera， O. (2015)Multiscale influences on extremewinter rainfall in the Philippines. J. Geophys. Res-Atmos..doi：10.1002/2014JD022645. Roberts， J.J.， Best， B.D.， Dunn， D.C.， Treml， E.A. and Halpin，P.N.(2010) Marine Geospatial Ecology Tools：；： Anintegrated framework for ecological geoprocessing withArcGIS， Python， R， MATLAB， and C++. Environ. Model.Softw. 25：1197-1207. Rooker， J.R.， Secor， D.H.， Zdanowicz， V.S.， De Metrio， G. andRelini， L.O. (2003) Identification of Atlantic bluefin tuna(Thunnus thynnus) stocks from putative nurseries usingotolith chemistry. Fish Oceanogr. 12：75-84. ( Rooker， J.R.， Secor， D.H.， De Metrio， G.， Kaufman，A.J.， Rios，A.B. and T icina， V . ( 2008a) Evidence o f trans-Atlanticmovement a n d n a tal h o ming of bluefin tuna from stableisotopes i n otoliths. Mar. E col. Prog. Ser. 368：231-2 39 . ) ( R ooker， J . R.， Secor， D .H.， D e Metrio， G . ， S c hloesser， R . ， Block， B.A. and Neilson， J.D. (2008b) Natal homing an connectivi t y in A t lantic bluefin tu n a populations. Science 322：742-744. ) ( Rooker， J.R.， Arrizabalaga， H.， Fr a ile， I. e t al. ( 2014) C r ossingthe line： migratory and homing b e haviors of Atlantic bluefintuna. Mar. E col. P rog. Ser. 504：265-276. ) ( Schaefer， K .M. ( 2 008) Stock structure of bigeye， yellowfin， andskipjack t unas i n the e astern P acific Ocean. In： I nter-Am. T rop. Tuna Comm. Stock Assess. Rep. 9：203- 2 21. ) Schaefer， K.M. and Fuller， D.W. (2009) Horizontal movementsof bigeye tuna (Thunnus obesus) in the eastern Pacific Ocean，as determined from conventional and archival taggingexperiments initiated during 2000-2005. Inter-Am. Trop.Tuna Comm. Bull. 24：191-247. Schaefer， K.M.， Fuller， D.W. anddMiyabe， N. ((2005)Reproductive biologhry of bigeye tuna (Thunnus obesus) in theeastern and central Pacific Ocean. Inter-Am. Trop. TunaComm. Bull. 23：1-31. Schaefer， K.M.， Fuller， D.W. anddBlock，， B.A. (2011)Movements， behaviour， and habitat utilization of yellowfintuna (Thunnus albacares) in the Pacific Ocean off BajaCalifornia， Mexico， determined from archival tag dataanalysis， including unscented Kalman filtering. Fish. Res.112：22-37. Schaefer， K.M.， Fuller， D.W.， Hampton， J.， Caillot， S.， Leroy， B.and Itano， D. (2015) Movements， dispersion， and mixing ofbigeye tuna (Thunnus obesus) tagged and released in theequatorial Central Pacific Ocean， with conventional andarchival tags. Fish. Res. 161：336-355. Shiao， J.C.， Wang， S.W.， Yokawa， K. et al. (2010) Natal originof Pacific bluefin tuna Thunnus orientalis inferred from otolithoxygen isotope composition. Mar. Ecol. Prog. Ser. 420：207-219. ( Sibert， J. and Hampton，J.( 2 003) Mobility of tropical tunas and t he implications for fisheries management. M a r. Policy27：87-95. ) ( Sturgeon， R.E.， Willie， S.N.， Yang， L. et a l . (2005) Certificationof a f ish otolith reference material in support of q uality assurance for t race element analysis. J. Anal. Atom. Spectrom. 20：1067 - 107 1 . ) ( Thorrold， S.R.， Campana， S .E.， Jones， C.M. and S wart， P .K.(1997) F actors determining 8 13 C and 8 1 8 Ofractionationin a ragonitic o toliths o f m arine f i sh. G eochim. C o smochim.Ac. 61：2909-2919. ) Thorrold， S.R.， Latkoczy， C.， Swart， P.K. and Jones， C.M.(2001) Natal homing in a marine fish metapopulation.Science 291：297-299. Walther， B.D.， Thorrold， S.R. and Olney， J.E. (2008)Geochemical signatures in otoliths record natal origins ofAmerican shad. Trans. Am. Fish. Soc. 137：57-69. ( Ward， R.D. ， E l liot， N.G.， I nnes， B.H.， S molenski， A .J. a nd Grewe， P . M. (1997) Globa l population str u cture of yellowfin tuna Thunnus albacares， j i nferred from allozyme andmitochondrial DNA variation. Fish. Bull. 95：566-575. ) Wells，R.J.D.， Rooker， J.R...anddPrince.E.D. ((2010)Regional1 variation iinn tthhee otolith chemistry ofbluemarlin (Makaira nigricans) and white marlin (Tetrapturusalbidus) from the western North Atlantic Ocean. Fish.Res. 106：430-435. ( Wells， R.J.D . ， Rooker， J.R. and Itano， D.G. (2012) N ursery origin of yellowfin tuna in the Hawaiian I slands. Mar. Ecol. Prog. Ser. 461：187-196. ) ( Y oshinaga， J.， Nakama， A .， M orita， M . and E d monds， J. S .(2000) Fish o tolith reference material for quality assurance of chemical analyses. Mar. Chem. 69：91-97. ) ( Y oung， J .W.， L ansdel l ， M.J.， Campbell， R.A.， C ooper， S .P .， J uanes， F. and G uest， M .A. ( 2 010) F e eding ecology and niche s egregation in o ceanic top p redators off e eastern Australia. Mar. Biol . 157：2347 - 2368. ) @ John Wiley & Sons Ltddoi：fog. C John Wiley & Sons Ltd， Fish. Oceanogr.， 大眼金枪鱼和黄鳍金枪鱼鱼耳石中的天然化学特征（稳定同位素和微量元素特征）被用来研究它们在西太平洋和中太平洋的起源以及分布情况。来自西太平洋和中太平洋四个区域的幼年 (YOY) T. obesus和 T. albacares 的鱼耳石的化学特征被首次研究并且用这些数据作为四个区域的背景值。幼年 T. obesus 和 T. albacares 的鱼耳石中稳定同位素的空间差异被研究，最明显的差异是这两种鱼类的耳石δ18O值相对于赤道中部和夏威夷地区，来自遥远的赤道西部和西赤道地区负偏。幼年 T. obesus 和 T. albacares 的鱼耳石元素比值在2008年被测定，一些数据有望区分YOY T. obesus(Mg:Ca, Mn:Ca, and Ba:Ca) 和T.albacares (Li:Caand Sr:Ca)两种耳石。1岁和2岁的T. obesus 和 T. albacares 在西太平洋和中太平洋的产地被鉴定出来。混合种群分析表明，西赤道样品中的T.obesus和T.albacares几乎完全来自当地的生产，很少来自赤道中部水域。同样，在夏威夷采集到的T.albacares也完全来自当地。然而，很大一部分夏威夷T. obesus 被划分为赤道中部地区，这表明来自产区以外区域（即夏威夷南部）的种群迁移对国内渔业很重要。